Any ideas about accelerating Croquet in-world dynamics with physics co-processors?

Now that physics co-processors are becoming available as COTS hardware, I
was wondering if folks are thinking about how to exploit them for Croquet?

See http://www.reghardware.co.uk/2006/03/29/asus_to_offer_ageia_physx/

   Best,

      Ed

The first question is whether the API allows for a repeatable  
experience.  We can't have the computations be different on different  
machines (e.g., because the random numbers are seeded differently or  
produce non-repeatable results, or if there are floating point  
computations that produce different results on different boards).

On Mar 29, 2006, at 1:12 PM, Ed Boyce wrote: